Post by thread:Buad timing- How close do I have to be?

Hi everyone I hace a project where a stamp BS2 and a Pic16F84 are talking back and forth on the same board (about an inch apart). The code utilizes 9600 baud. I have a couple of questions. 1) For 9600 baud and bit timing should be 104us. The best I have been able to do in my 16F84 program is 116us between bits, due to the way I have implemented the code. This is consistent for both the serout and serin routines. Can I communicate effectively with the BS2 with the slight offset in timing? The BS2 will utilize its standard serin/out commands. 2) What kind of limitations do I have for baud rate in this application? Can I run at 19,200 or even higher? Thanks in advance Jon

On Sun, 24 May 1998 12:46:23 EDT PHXSYS <spam_OUTPHXSYSTakeThisOuTAOL.COM> writes: >1) For 9600 baud and bit timing should be 104us. The best I have been >able to >do in my 16F84 program is 116us between bits, due to the way I have >implemented the code. This is consistent for both the serout and >serin >routines. Can I communicate effectively with the BS2 with the slight >offset in >timing? The BS2 will utilize its standard serin/out commands. You can have a slight difference, but 10% is too much. About 5% is the limit. Usually a PIC crystal can be selected so the baud rate works out exactly with your software approach. >2) What kind of limitations do I have for baud rate in this >application? Can I >run at 19,200 or even higher? There is no problem running high baud rates over a short TTL-level link. It will work as fast as your software can go. _____________________________________________________________________ You don't need to buy Internet access to use free Internet e-mail. Get completely free e-mail from Juno at http://www.juno.com Or call Juno at (800) 654-JUNO [654-5866]

PHXSYS wrote: > Hi everyone > > I hace a project where a stamp BS2 and a Pic16F84 are talking back and forth > on the same board (about an inch apart). The code utilizes 9600 baud. I have a > couple of questions. > > 1) For 9600 baud and bit timing should be 104us. The best I have been able to > do in my 16F84 program is 116us between bits, due to the way I have > implemented the code. This is consistent for both the serout and serin > routines. Can I communicate effectively with the BS2 with the slight offset in > timing? The BS2 will utilize its standard serin/out commands. > > 2) What kind of limitations do I have for baud rate in this application? Can I > run at 19,200 or even higher? Start off with a crystal that will give you zero error. They are available closeto most 'popular' frequencies. 1.8432MHz, 3.6864MHz, and 11.0592MHz spring to mind. You can get all the standard baud rates with these crystals. I can't remember the other frequencies, but I'm sure someone would add to this message. -- Friendly Regards Tjaart van der Walt .....tjaartKILLspam@spam@wasp.co.za |--------------------------------------------------| | WASP International | |R&D Engineer : GSM peripheral services development| |--------------------------------------------------| |SMS 0832123443KILLspamwasp.co.za (160 chars max)| | http://www.wasp.co.za/~tjaart/index.html | |Voice: +27-(0)11-622-8686 Fax: +27-(0)11-622-8973| | WGS-84 : 26¡10.52'S 28¡06.19'E | |--------------------------------------------------|

On Mon, 25 May 1998 09:44:04 -0600 Matt Bonner <.....mbonnerKILLspam.....SUNADA.COM> writes: >Tjaart van der Walt wrote: >> 1.8432MHz > 2.4576MHz ... Any more? >> 3.6864MHz >> 11.0592MHz > >> You can get all the standard baud rates with these crystals. There are two major 'series' of frequencies, with standard crystals available at each power of 2 in the series: (9600 * 192): 1.8432, 3.6864, 7.3728, 14.7456 MHz (9600 * 256): 2.4576, 4.9152, 9.8304, 19.6608 MHz A 9.8304 is a good match for a 10 MHz PIC16F84, or 19.6608 with a PIC rated for 20 MHz. Most 4 MHz PICs have no problem with 4.9152, but use 3.6864 to be safe. Also fairly common are 18.432 (1.8432 * 10) and 11.0592 (1.8432 * 6). The latter is very popular with 8051 type controllers because of the way they divide the clock. Frequencies of 6.144 (9600 * 640) and 12.288 MHz were commonly used with Intel parts rated for 6 or 12 MHz respectively. _____________________________________________________________________ You don't need to buy Internet access to use free Internet e-mail. Get completely free e-mail from Juno at http://www.juno.com Or call Juno at (800) 654-JUNO [654-5866]

At 09:44 25/05/98 -0600, you wrote: >Tjaart van der Walt wrote: >> 1.8432MHz > 2.4576MHz ... Any more? >> 3.6864MHz >> 11.0592MHz > >> You can get all the standard baud rates with these crystals. >> >> I can't remember the other frequencies, but I'm sure someone would >> add to this message. 7.3728 9.2160 etc (I'm using these 2) Basically if it's correct (0% error) for 1.8432MHz (or for a lower Xtal frequency) then (because the baud rate divisor is a power of 2) then take any multiple you like! (Except I get 2.7648 instead of the above 3.4576) The above is for BRGH=0, modify accordingly (factor of 16) for BRGH=1 (or whatever). Regards Adrian WWW WWW Adrian Gothard WWW WW WWW White Horse Design WWWWWWWWWW +44-385-970009 (Mobile/SMS), +44-118-962-8913/4 (voice/fax) WWWW WWWW EraseMEwhdspam_OUTTakeThisOuTzetnet.co.uk, http://www.users.zetnet.co.uk/whd --- Developers of GPS satellite-based tracking systems

At 12:46 24/05/98 EDT, you wrote: >Hi everyone > >I hace a project where a stamp BS2 and a Pic16F84 are talking back and forth >on the same board (about an inch apart). The code utilizes 9600 baud. I have a >couple of questions. > >1) For 9600 baud and bit timing should be 104us. The best I have been able to >do in my 16F84 program is 116us between bits, due to the way I have >implemented the code. This is consistent for both the serout and serin >routines. Can I communicate effectively with the BS2 with the slight offset in >timing? The BS2 will utilize its standard serin/out commands. > >2) What kind of limitations do I have for baud rate in this application? Can I >run at 19,200 or even higher? I just replied to your later message. If you're doing this in software (or hardware for that matter) I always aim for an error of less than 1%, and I try to get the best I possibly can especially when using equipment with different UARTs or unknown UARTs. For a software implementation, check for the start bit periodically (or have it interrupt driven) then start a (software or hardware) timer for half the bit time. Resample the start bit if valid, start a counter for one bit time (then you'll be in the middle of bit 0 (lsb comes first of course), repeat 5/6/7/8 times (or 1 more than each of these if you have a parity bit). As a "slight cheat" I sample about 1/4 after the bit transition time since I am using a software implementation with various polling delays within the start bite period time. Regards Adrian WWW WWW Adrian Gothard WWW WW WWW White Horse Design WWWWWWWWWW +44-385-970009 (Mobile/SMS), +44-118-962-8913/4 (voice/fax) WWWW WWWW whdspam_OUTzetnet.co.uk, http://www.users.zetnet.co.uk/whd --- Developers of GPS satellite-based tracking systems

Karsten Krause wrote: > > Hi Matt, > > please excuse my writing to you, but I just subscribed to the piclist. > I subscribed for the digest version. Since all my emails are answered > by their machine :( - do you have any ideas if there is such an option > as digest possible ? > Since I seem to be the only Matt who has contributed to the baud timing thread, I'll assume this is addressed to me :-) The following is from an Andy Warren post this morning (easier than going into my 'Lists' folder where I keep _all_ my list subscribe information - don't we all?): To subscribe to the PICLIST, send a message to: @spam@listservKILLspammitvma.mit.edu with the following in the BODY of the message (the message subject is unimportant): subscribe piclist set piclist repro end Once you're subscribed, you can switch to digest mode by sending the following to the same address: set piclist digest end --Matt

Mike Keitz wrote: > > On Mon, 25 May 1998 09:44:04 -0600 Matt Bonner <KILLspammbonnerKILLspamSUNADA.COM> > writes: > >Tjaart van der Walt wrote: > >> 1.8432MHz > > 2.4576MHz ... Any more? > >> 3.6864MHz > >> 11.0592MHz > > > >> You can get all the standard baud rates with these crystals. > > There are two major 'series' of frequencies, with standard crystals > available at each power of 2 in the series: > > (9600 * 192): 1.8432, 3.6864, 7.3728, 14.7456 MHz > (9600 * 256): 2.4576, 4.9152, 9.8304, 19.6608 MHz Also, you can buy 3.579545MHz and 14.318MHz crystals -- these are based on the NTSC colour burst frequency, and due to their appearance in colour TVs (especially the former) they are very cheap. For the 3.58MHz crystal you need a bit loop of 93 instructions, with an accuracy of better than 0.07%. Frank ------------------------------------------------------------------------ Frank A. Vorstenbosch <UCE_ACCEPT="NONE"> Mobile: +44-976-430 569 Wimbledon, London SW19 Home: +44-181-544 1865 RemoveMEfrankTakeThisOuTfalstaff.demon.co.uk Office: +44-181-636 3391

Mike, to follow-up, I've been looking into crystals as part of my PIC-based logic analyzer. For folks using the 16C76/77 parts, the following common crystals will provide 115.2Kbaud and 0% error with BRGH = 1: 3.6864 Mhz -> SPBRG = 1 7.3728 Mhz -> SPBRG = 3 11.0592 Mhz -> SPBRG = 5 14.7456 Mhz -> SPBRG = 7 18.4320 Mhz -> SPBRG = 9 - Tom At 12:39 PM 5/25/98 -0400, Mike Keitz wrote: {Quote hidden}>On Mon, 25 May 1998 09:44:04 -0600 Matt Bonner <spamBeGonembonnerspamBeGoneSUNADA.COM> >writes: >>Tjaart van der Walt wrote: >>> 1.8432MHz >> 2.4576MHz ... Any more? >>> 3.6864MHz >>> 11.0592MHz >> >>> You can get all the standard baud rates with these crystals. > >There are two major 'series' of frequencies, with standard crystals >available at each power of 2 in the series: > >(9600 * 192): 1.8432, 3.6864, 7.3728, 14.7456 MHz >(9600 * 256): 2.4576, 4.9152, 9.8304, 19.6608 MHz > >A 9.8304 is a good match for a 10 MHz PIC16F84, or 19.6608 with a PIC >rated for 20 MHz. Most 4 MHz PICs have no problem with 4.9152, but use >3.6864 to be safe. Also fairly common are 18.432 (1.8432 * 10) and >11.0592 (1.8432 * 6). The latter is very popular with 8051 type >controllers because of the way they divide the clock. Frequencies of >6.144 (9600 * 640) and 12.288 MHz were commonly used with Intel parts >rated for 6 or 12 MHz respectively. > > >_____________________________________________________________________ >You don't need to buy Internet access to use free Internet e-mail. >Get completely free e-mail from Juno at http://www.juno.com >Or call Juno at (800) 654-JUNO [654-5866] > >

Frank A. Vorstenbosch wrote: {Quote hidden}> Mike Keitz wrote: > > > > On Mon, 25 May 1998 09:44:04 -0600 Matt Bonner <TakeThisOuTmbonnerEraseMEspam_OUTSUNADA.COM> > > writes: > > >Tjaart van der Walt wrote: > > >> 1.8432MHz > > > 2.4576MHz ... Any more? > > >> 3.6864MHz > > >> 11.0592MHz > > > > > >> You can get all the standard baud rates with these crystals. > > > > There are two major 'series' of frequencies, with standard crystals > > available at each power of 2 in the series: > > > > (9600 * 192): 1.8432, 3.6864, 7.3728, 14.7456 MHz > > (9600 * 256): 2.4576, 4.9152, 9.8304, 19.6608 MHz > > Also, you can buy 3.579545MHz and 14.318MHz crystals -- these are based on the > NTSC colour burst frequency, and due to their appearance in colour TVs > (especially the former) they are very cheap. True. If you calculate BRGH for 3.6864MHz for 2400 baud or 9600 baud,you can get away with the 3.579545MHz crystal if you can't get any 3.6's. -- Friendly Regards Tjaart van der Walt RemoveMEtjaartTakeThisOuTwasp.co.za |--------------------------------------------------| | WASP International | |R&D Engineer : GSM peripheral services development| |--------------------------------------------------| |SMS 0832123443EraseME.....wasp.co.za (160 chars max)| | http://www.wasp.co.za/~tjaart/index.html | |Voice: +27-(0)11-622-8686 Fax: +27-(0)11-622-8973| | WGS-84 : 26¡10.52'S 28¡06.19'E | |--------------------------------------------------|